Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 7 Sep 2018 21:49:31 -0700
From:      Steve Kargl <sgk@troutmask.apl.washington.edu>
To:        Bruce Evans <brde@optusnet.com.au>
Cc:        freebsd-numerics@freebsd.org
Subject:   Re: j0 (and y0) in the range 2 <= x < (p/2)*log(2)
Message-ID:  <20180908044931.GA52882@troutmask.apl.washington.edu>
In-Reply-To: <20180907175406.GA49351@troutmask.apl.washington.edu>
References:  <20180903235724.GA95333@troutmask.apl.washington.edu> <20180905201540.D1142@besplex.bde.org> <20180905152104.GA26453@troutmask.apl.washington.edu> <20180906034525.A2959@besplex.bde.org> <20180905190223.GA27865@troutmask.apl.washington.edu> <20180906202658.K978@besplex.bde.org> <20180906223026.GA43005@troutmask.apl.washington.edu> <20180907230825.A933@besplex.bde.org> <20180907175406.GA49351@troutmask.apl.washington.edu>

next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, Sep 07, 2018 at 10:54:06AM -0700, Steve Kargl wrote:
> 
> I'll probably try a different approach tomorrow.  j0(x),
> cc, and ss are smoothly varying functions of x.  u and v
> appear to be slowly varying (in small intervals), so we can
> have
> 
> j0(x)   = sqrt(1/pi) * (cc(x)  *u-ss(x)  *v) / sqrt(x).
> j0(x+e) = sqrt(1/pi) * (cc(x+e)*u-ss(x+e)*v) / sqrt(x+e).
>  
> where e is some small perturbution.  Two equation and two
> unknowns is easily solved.  Simply need to do this across
> the interval to generate u(x) and v(x).
> 

I may have have a better approach!

j0(x) = sqrt(1/pi) * (cc(x) * u - ss(x) * v) / sqrt(x)
y0(x) = sqrt(1/pi) * (cc(x) * u + ss(x) * v) / sqrt(x)

j0(x) + y0(x) =   2 * sqrt(1/pi) * cc(x) * u
j0(x) - y0(x) = - 2 * sqrt(1/pi) * ss(x) * v

So, we have

u = (j0 + y0) / (2 * sqrt(1/pi) * cc)
v = (y0 - j0) / (s * sqrt(1/pi) * ss)

Thus, we can find u = r/s or u = 1 + r/s and v = r/s.

-- 
Steve



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20180908044931.GA52882>